torch-geometric
Graph Neural Network Library for PyTorch
Decision gist · record as of 2026-08-14
Yes. PyTorch Geometric is production-stable (Development Status 5), actively maintained, has no known vulnerabilities, and is in the top 5000 PyPI packages by download volume. Install friction is low and the MIT license is permissive. It is the standard library for GNN work in PyTorch—install it if you need to build or train graph neural networks.AI-flagged interpretation of the facts on this page — verify before relying
Before you install
- PyTorch must be installed separately (not listed as a direct dependency in the fact sheet, but is the core runtime requirement).
- Requires Python >=3.10.
- Low friction install with a pure Python wheel.
License · maintenance · safety
MIT (permissive) — MIT license (permissive). You can use, modify, and distribute this package freely in commercial and private projects with minimal restrictions—just include the license notice.
last release 2026-07-20 (25 days) · last repo commit 2026-07-31 · 24,008 stars
0 known vulnerabilities (OSV.dev, 2026-08-14) · 1,517,719 downloads/mo, #3,821 on PyPI
Alternatives
Verify before relying
pip install torch-geometric
import torch
from torch_geometric.nn import GCNConv
from torch_geometric.datasets import Planetoid
dataset = Planetoid(root='.', name='Cora')
model = GCNConv(dataset.num_features, dataset.num_classes)- Whether PyTorch is declared as a dependency in the actual package metadata (not shown in the runtime list).
- GPU/CUDA support requirements and whether optional compiled extensions (pyg-lib) are automatically built or require system libraries.
What it is and what it does
PyTorch Geometric is a framework for implementing Graph Neural Networks on top of PyTorch. It provides a collection of pre-built GNN layers (GCNConv, GraphSAGE, GAT, etc.), a message-passing API for custom architectures, mini-batch loaders for both small and large graphs, and benchmark datasets. The library is designed to feel like native PyTorch—if you know PyTorch, the API is straightforward.
You use it to solve problems on graph-structured data: node classification, link prediction, graph classification, and other geometric deep learning tasks. It handles diverse graph types including static graphs, dynamic graphs, heterogeneous graphs with multiple node and edge types, and 3D point clouds. The library supports multi-GPU training and torch.compile for performance optimization.
Use it for
- Classify nodes in citation networks or social graphs using pre-built GCN or GraphSAGE layers.
- Build custom GNN architectures by extending the MessagePassing base class for research or domain-specific problems.
- Train on large-scale graphs with millions of nodes using mini-batch loaders and scalable GNN models.
- Perform link prediction or graph-level classification on benchmark datasets (OGB, Cora, Citeseer, etc.).
- Process 3D point cloud data or mesh structures using geometric transformations and specialized layers.
- Prototype heterogeneous graph models with multiple node and edge types for knowledge graphs or recommendation systems.
Worth the install?
AI-flagged interpretation of the facts on this page. Verify before relying on it.
Yes.
PyTorch Geometric is production-stable (Development Status 5), actively maintained, has no known vulnerabilities, and is in the top 5000 PyPI packages by download volume. Install friction is low and the MIT license is permissive. It is the standard library for GNN work in PyTorch—install it if you need to build or train graph neural networks.
Install
torch-geometric on PyPI
Before you install
Low friction install with a pure Python wheel. Active maintenance—last commit 2026-07-31, release 25 days ago. Supports Python 3.10 through 3.14. Nine runtime dependencies are all well-established packages (numpy, torch, aiohttp, requests, tqdm, etc.), so no unusual compatibility risks.
PyTorch must be installed separately (not listed as a direct dependency in the fact sheet, but is the core runtime requirement). Requires Python >=3.10.
License in practice
MIT license (permissive). You can use, modify, and distribute this package freely in commercial and private projects with minimal restrictions—just include the license notice.
Quickstart
pip install torch-geometric
import torch
from torch_geometric.nn import GCNConv
from torch_geometric.datasets import Planetoid
dataset = Planetoid(root='.', name='Cora')
model = GCNConv(dataset.num_features, dataset.num_classes)Verify before relying
- Whether PyTorch is declared as a dependency in the actual package metadata (not shown in the runtime list).
- GPU/CUDA support requirements and whether optional compiled extensions (pyg-lib) are automatically built or require system libraries.
Package facts
| License | MIT permissive |
| Python support | Supports the current Python release >=3.10 |
| Install friction | Low. Pure-Python wheel |
| Runtime dependencies | 9 packagesaiohttpfsspecjinja2numpypsutilpyparsingrequeststqdmxxhash |
| Maintenance | Actively maintained 25 days since the last release |
| Last repo commit | |
| First released | |
| Downloads | 1,517,719 / month, #3,821 on PyPI 30-day window, as of 2026-08-14 |
| Known vulnerabilities | None known OSV.dev, checked 2026-08-14 |
| Classifiers | Development Status :: 5 - Production/StableProgramming Language :: PythonProgramming Language :: Python :: 3 :: OnlyProgramming Language :: Python :: 3.10Programming Language :: Python :: 3.11Programming Language :: Python :: 3.12Programming Language :: Python :: 3.13Programming Language :: Python :: 3.14 |
Evidence: torch_geometric-2.8.0.post1-py3-none-any.whl
Tags
Let your AI agent find packages like this
Example. Real query, live index.
You found this page by searching. An agent finds it by wishing: SkillFed indexes 14,416 PyPI packages by what they can do, searchable in plain language.
wish › “GNN library pytorch”
- torch-geometricPyTorch Geometric is a library for building and training Graph Neural…
- ogbOGB provides standardized benchmark datasets, data loaders, and…
- jraphJraph provides data structures and utilities for building and working…
Give your agent the search over MCP, or paste the wish link into any chat.
More Artificial Intelligence packages
LiteLLM provides a unified Python interface to call 100+ LLM providers (OpenAI, Anthropic, Gemini, Bedrock, Azure, and others) using OpenAI-compatible API format, available as both a Python SDK and a self-hosted AI Gateway proxy server.
Install it if you need to work with multiple LLM providers or want to centralize LLM routing in your organization.
Client library and CLI tool for downloading, uploading, and managing models, datasets, and repositories on the Hugging Face Hub platform.
Install it if you work with Hugging Face Hub models or datasets.
LangChain provides a framework for building agents and LLM-powered applications by composing language models, tools, and memory through a unified API that abstracts over multiple model providers.
hf-xet provides chunk-based deduplication and efficient file transfer for the Hugging Face Hub, enabling faster uploads and downloads of large files with local disk caching.
Tokenizers converts raw text into token sequences for NLP models, with support for training custom vocabularies and using pre-built tokenizers (BPE, WordPiece) optimized for speed via Rust.
Transformers provides a unified framework for loading, fine-tuning, and running state-of-the-art pretrained models across text, vision, audio, video, and multimodal tasks using PyTorch, JAX, or TensorFlow.
Install it if you need to run or train any transformer-based model for NLP, vision, audio, or multimodal tasks.
See also torch · ogb · pytorchcv · cuequivariance-torch · e3nn · jraph · torchinfo · pnnx · torch-complex · segmentation-models-pytorch